Unraveling recombination rate evolution using ancestral recombination maps

Abstract

Recombination maps of ancestral species can be constructed from comparative analyses of genomes from closely related species, exemplified by a recently published map of the human-chimpanzee ancestor. Such maps resolve differences in recombination rate between species into changes along individual branches in the speciation tree, and allow identification of associated changes in the genomic sequences. We describe how coalescent hidden Markov models are able to call individual recombination events in ancestral species through inference of incomplete lineage sorting along a genomic alignment. In the great apes, speciation events are sufficiently close in time that a map can be inferred for the ancestral species at each internal branch - allowing evolution of recombination rate to be tracked over evolutionary time scales from speciation event to speciation event. We see this approach as a way of characterizing the evolution of recombination rate and the genomic properties that influence it.

Keywords: evolution; genomics; recombination.

Figure 1

Incomplete lineage sorting reveals ancestral recombination events. A: Species tree for three species with lines at the tips representing sequences. A crossover (red dot) in the ancestral species lineage defines two subsequences (orange and blue) with separate genealogies. B: Alignment of the three sequences from panel A with underlying genealogies. The red dashed line represents the position of the recombination event.

Figure 2

Inference of ancestral recombination events. A: The four hidden states of the coalescent HMM which alternate along the alignment. B: Posterior probabilities of each hidden state along 3 kb of alignment. The sequence of most likely genealogies is shown at the top.

Figure 3

A sequence of hypothetical ancestral recombination maps. Colored dots on the cladogram of old-world monkeys (center) each marks the ancestral species represented by each internal branch. Arrows from each dot point to the species tree of three extant species whose genomes can be used to produce a map of that ancestral species. The color intensity shows the expected distribution of recombination events measured by the constructed recombination map. Numbers represent the speciation times in millions of years used in simulations. Effective population sizes of ancestral species used are: bonobo-chimpanzee: 32,000, human-chimpanzee: 86,000, human-gorilla: 62,000, human-orangutan: 167,000, human-gibbon: 167,000.

Figure 4

Number of recombination events included in each of the five ancestral recombination maps in Fig. 3 and their distribution in time. Dashed lines represent speciation times.

Figure 5

Species tree as in Fig. 3 of Neanderthal, Denisovan, and modern human. Color intensity shows the expected distribution recombination events measured by this hypothetical recombination map. Numbers represent speciation times in years used in simulations. All effective population sizes are set to 10,000.